RPO and backup


After reading some topics here, i get a question about backup and RPO.

If I take a snapshot every day, my RPO could be 24h : that's a lots.

But taking a snapshot every hour or every 10' or more is a lots too ...

With Elasticsearch is there something like Oracle transaction logs ?

I remembered that there are some luccene buffers but they are in memory and as i understood they can not be used to restore a cluster.

Thanks for any help.

B. Granier
Elasticsearch certified

You could use cross cluster replication to replicate to a separate cluster. That is probably the best way to minimize RPO.

Yes but for that a "new" cluster is needed and there is Always the question of the distance between clusters nodes.

A new cluster is needed but distance should not be a problem with cross cluster replication as it is designed to replicate data across regions.